摘要 隨著科技的發展與進步,人類生活逐漸改善與便利化,然而,老化一直是人類目前尚無法克服的問題。人類於膝蓋有一塊半月軟骨,會隨著不停的使用與重量的壓迫漸漸的被磨耗掉。下肢輔具發展至今已經有多年的歷史,本實驗室也發展了數代下肢輔具,本文也將延續過去研究並改進,讓下肢輔具更完善與實用。 肌電訊號(Electromyography, 或EMG)是人體肌肉在實際收縮前所會發出的微弱電子訊號,至今有許多研究針對肌電訊號與肌肉做動之間一些關聯與特性,透過了解其特性,本文輔具以肌電訊號為主要判斷依據,藉由ANFIS模型的訓練,使得輔具能夠即時的透過肌電訊號分析控制力並適時、適量地給予輔助。研究顯示肌肉之動作還可分為離心收縮與向心收縮,兩者的振幅力矩之對應是不同的,本文將修正此問題,使輔具能夠更精確地輔助人體。 輔具硬體以特殊設計之M型彈簧為主要的儲能部件,動態地將彈簧做拉伸與釋放的動作,藉由從動輔助模式、支撐模式與擺盪模式動態的切換,在膝蓋需求力矩的時候釋放彈簧給予與身體輔助,不需求額外力矩的時候讓身體不感受到輔具的阻力,相信在此適時適量的輔助,可以改善膝蓋軟骨的磨耗並提升年長者的行動距離。
Abstract The advancing of technology can enhance human life. However, aging is still a problem for human. Through time, Meniscus will gradually wear out with continuous loading. Lower-limb exoskeleton had been studied for several years; this thesis follows previous research and focus on perfection in lower-limb exoskeleton. Electromyography (EMG) will transmit through muscle while muscle contraction. Researches had pointed out relationship between muscle activity and EMG. By knowing these characteristics, our lower-limb exoskeleton can be trained through an Adaptive Network-based Fuzzy Interface System (ANFIS), to control our exoskeleton instantaneously. Other researchers pointed out that muscle contraction can be divided into two types, concentric and eccentric. With the difference in muscle contraction, the model of EMG and moment should be adjusted. Accordingly, this thesis will deal with this problem to create a better principle in exoskeleton control. Our exoskeleton is based on special designed M shaped springs, which were used as energy storage element. We dynamically change the system mode to stance mode, swing mode or non-effect mode, with different supporting way in different modes, our exoskeleton pulls and releases the spring dynamically to store and provides the potential energy to our body while in need. With this idea of assisting in the correct time and the correct amount of energy, we would be able to ease off the problem of deteriorating cartilage and prolong the ability for elders to travel.